The formation of complex households among immigrants is an important strategy for adapting to the migration process and adjusting to a new society. Complex households often represent a conduit for accessing information about services and programs, provide a context for the delivery and receipt of care, form an environment for maintaining cultural identity, and offer a platform for sharing economic resources. The purpose of this project is to examine the living arrangements of older immigrants. The research will be guided by our extension of the resource-preference model of living arrangement behavior. This project focuses on Hispanic and Asian immigrants, comparing these groups to native-born Hispanics, Asians, Whites and African Americans. The primary data sources are the 1990 and 2000 Censuses of Population. Several questions are addressed. What are the patterns of household composition across immigrant groups and what individual factors are associated with these patterns? What is the impact of English language ability and duration of residence on living arrangement patterns? How does the ethnic dimension of communities condition the likelihood of household complexity? Are some immigrant groups more likely than others to form complex versus independent households when living in the context of an ethnic community? How have older immigrant's living arrangements changed during the 1990s? With respect to changes in living arrangements, what roles do life cycle stage and assimilation explanations play in understanding the adjustment process? Three specific aims are offered to address these questions. First, we will compare the health, socio-demographic and economic characteristics of persons who co-reside with older immigrants with the immigrants themselves to determine patterns of family relationships and dependency among household members. An innovative household relationship indexing strategy is employed to overcome limitations of household relationship indicators provided in Census data. Second, multilevel models of living arrangement behavior among specific groups of older immigrants and non-immigrants will be estimated to focus on how language ability, ethnic community characteristics, and other spatial characteristics work independently and jointly to influence this behavior. Hierarchical non-linear modeling techniques are employed. Third, using the double cohort method, age and immigration cohort effects on patterns of household complexity are examined to test life cycle and assimilation hypotheses.